TY - JOUR
T1 - Fully automated deep learning based auto-contouring of liver segments and spleen on contrast-enhanced CT images
AU - Gupta, Aashish C.
AU - Cazoulat, Guillaume
AU - Al Taie, Mais
AU - Yedururi, Sireesha
AU - Rigaud, Bastien
AU - Castelo, Austin
AU - Wood, John
AU - Yu, Cenji
AU - O’Connor, Caleb
AU - Salem, Usama
AU - Silva, Jessica Albuquerque Marques
AU - Jones, Aaron Kyle
AU - McCulloch, Molly
AU - Odisio, Bruno C.
AU - Koay, Eugene J.
AU - Brock, Kristy K.
N1 - © 2024. The Author(s).
PY - 2024/2/26
Y1 - 2024/2/26
N2 - Manual delineation of liver segments on computed tomography (CT) images for primary/secondary liver cancer (LC) patients is time-intensive and prone to inter/intra-observer variability. Therefore, we developed a deep-learning-based model to auto-contour liver segments and spleen on contrast-enhanced CT (CECT) images. We trained two models using 3d patch-based attention U-Net (MpaU-Net) and 3d full resolution of nnU-Net (MnnU-Net) to determine the best architecture (BA). BA was used with vessels (MVess) and spleen (Mseg+spleen) to assess the impact on segment contouring. Models were trained, validated, and tested on 160 (CRTTrain), 40 (CRTVal), 33 (CLS), 25 (CCH) and 20 (CPVE) CECT of LC patients. MnnU-Net outperformed MpaU-Net across all segments with median differences in Dice similarity coefficients (DSC) ranging 0.03–0.05 (p < 0.05). Mseg+spleen, and MnnU-Net were not statistically different (p > 0.05), however, both were slightly better than MVess by DSC up to 0.02. The final model, Mseg+spleen, showed a mean DSC of 0.89, 0.82, 0.88, 0.87, 0.96, and 0.95 for segments 1, 2, 3, 4, 5–8, and spleen, respectively on entire test sets. Qualitatively, more than 85% of cases showed a Likert score ≥ 3 on test sets. Our final model provides clinically acceptable contours of liver segments and spleen which are usable in treatment planning.
AB - Manual delineation of liver segments on computed tomography (CT) images for primary/secondary liver cancer (LC) patients is time-intensive and prone to inter/intra-observer variability. Therefore, we developed a deep-learning-based model to auto-contour liver segments and spleen on contrast-enhanced CT (CECT) images. We trained two models using 3d patch-based attention U-Net (MpaU-Net) and 3d full resolution of nnU-Net (MnnU-Net) to determine the best architecture (BA). BA was used with vessels (MVess) and spleen (Mseg+spleen) to assess the impact on segment contouring. Models were trained, validated, and tested on 160 (CRTTrain), 40 (CRTVal), 33 (CLS), 25 (CCH) and 20 (CPVE) CECT of LC patients. MnnU-Net outperformed MpaU-Net across all segments with median differences in Dice similarity coefficients (DSC) ranging 0.03–0.05 (p < 0.05). Mseg+spleen, and MnnU-Net were not statistically different (p > 0.05), however, both were slightly better than MVess by DSC up to 0.02. The final model, Mseg+spleen, showed a mean DSC of 0.89, 0.82, 0.88, 0.87, 0.96, and 0.95 for segments 1, 2, 3, 4, 5–8, and spleen, respectively on entire test sets. Qualitatively, more than 85% of cases showed a Likert score ≥ 3 on test sets. Our final model provides clinically acceptable contours of liver segments and spleen which are usable in treatment planning.
KW - Humans
KW - Spleen/diagnostic imaging
KW - Deep Learning
KW - Tomography, X-Ray Computed/methods
KW - Liver Neoplasms/diagnostic imaging
KW - Image Processing, Computer-Assisted/methods
UR - http://www.scopus.com/inward/record.url?scp=85186218672&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85186218672&partnerID=8YFLogxK
U2 - 10.1038/s41598-024-53997-y
DO - 10.1038/s41598-024-53997-y
M3 - Article
C2 - 38409252
AN - SCOPUS:85186218672
SN - 2045-2322
VL - 14
SP - 4678
JO - Scientific Reports
JF - Scientific Reports
IS - 1
M1 - 4678
ER -